A passive radar anti-active decoy method based on DOA clustering algorithm

Abstract

The present invention provides a passive radar anti-active decoy method based on a DOA clustering algorithm, which determines whether there are multiple radiation source targets by the power and phase difference jump of the received signal, and performs DOA estimation; within a period of time, the angle results of the DOA estimation are accumulated and dynamic distance clustering is performed; the clustering results are judged to determine whether they meet the conditions for determining isolated points; the clustering results are calculated for repetition frequency stability, and the target angle is output. The present invention can accurately determine whether there is active decoy in the current environment, has high real-time performance, low algorithm complexity, and can realize the recognition and judgment of active decoy system targets. It is suitable for airborne / missile-borne passive radars to find and identify ground fixed-position radiation source targets, and belongs to a ground target precision guidance system, with high real-time requirements and low misjudgment rate.

Description

Technical Field

[0001] The invention relates to the technical field of measurement and testing, and in particular to a passive radar anti-active decoy method based on a DOA clustering algorithm. Background Art

[0002] Passive radar can detect and acquire signals from active non-cooperative targets (such as ground detection radars, communication stations, etc.). It is an extremely important means of target detection. It has strong anti-reconnaissance, anti-interference, anti-hard kill, stealth target and low-altitude and ultra-low-altitude penetration target detection capabilities, which makes up for the shortcomings of active radar. Since it does not carry an electromagnetic wave radiation source, it has a very strong anti-reconnaissance capability, and its detection capability for stealth targets is stronger than that of general active radars. Based on the angle estimation of the leading part of the arrival signal, the flashing bait and target information with different arrival times can be extracted. On the basis of obtaining the active decoy angle information, relatively isolated targets can be clustered and sorted out.

[0003] Xu Duan, Dong Wenfeng, Zhou Wu, Research on the performance of ARUAV front discrimination method against multi-point source decoy, Command Control and Simulation, 2012, Vol.34, No.4.4, pp.406-423. This paper proposes that the anti-radiation unmanned aerial vehicle (ARUAV) uses the front tracking technology based on wavelet transform to counter the multi-point source decoy system, effectively counter the two-point source decoy system, and obtains an effective discrimination airspace by the fixed wave gate. This method has high requirements on the duration of the single front signal. When the exposure time of the single signal is short, the effective discrimination airspace is small.

[0004] Dai Huanyao, Wang Jianlu, Li Deshen, A method for identifying radar and decoy interference by a passive radar seeker, 2020, CN202010736721.4. This patent proposes a method for determining the presence of decoy interference through the polarization-azimuth joint spectrum feature, and determining the target radar and decoy interference angle information using the polarization-azimuth joint spectrum estimation and interference identification algorithm. This method has high requirements on the system computing power, long multi-dimensional calculation time, and low real-time performance. Summary of the invention

[0005] The present invention is to solve the efficiency problem of passive radar anti-active decoy, and provides a passive radar anti-active decoy method based on DOA clustering algorithm, which uses the power and phase difference jump of the received signal to judge whether there are multiple radiation source targets, and uses real-time DOA estimation, DDC clustering and repetition frequency stability calculation to make the calculation result accurate. The present invention can accurately determine whether there is active decoy in the current environment, has high real-time performance, low algorithm complexity, can realize the recognition and judgment of active decoy system targets, and realizes the direction finding and recognition of fixed-position radiation source targets on the ground by airborne / missile-borne passive radars.

[0006] The present invention provides a passive radar anti-active decoy method based on a DOA clustering algorithm, comprising the following steps:

[0007] S1. Determine whether there are multiple radiation source targets based on the power and phase difference jump of the passive radar receiving signal. If so, perform DOA estimation to obtain the angle measurement result x i , angle measurement result x i for And go to step S2; if not, continue to receive signals;

[0008] S2. Accumulate the angle measurement results to obtain the angle measurement result sample set X and perform DDC clustering to obtain the clustering result. The clustering result is all classes, and all classes are z 1 ,z 2 ,L z q ;

[0009] S3, determine whether the center dispersion of the class is greater than the threshold, if yes, go to step S4, if not, return to step S1;

[0010] S4. Perform repetition frequency stability calculation on the clustering results to obtain the target angle.

[0011] The passive radar anti-active decoy method based on DOA clustering algorithm described in the present invention, as a preferred embodiment, step S1 includes the following steps:

[0012] S11. Calculate the passive radar two-way antenna receiving signal s x (t), s x+1 Phase difference (t) And the antenna receives the signal s x (t) and s x+1 (t) Sampling point data within the preset time threshold of the front part s x (m), s x+1 Phase difference (m) make and The difference is obtained by taking the phase difference ΔΦ, where x=1,...,X, x is initialized to 1, t=1, Λ, T; m=1, Λ, M; M=T;

[0013] S12, calculate the antenna receiving signal s x The average amplitude of (t) t , sampling point data s x The average amplitude of (m) m , A t With A m Taking the difference, we get the amplitude difference ΔA;

[0014] S13, judging whether the cumulative number of times that the phase difference ΔΦ exceeds 60° and the amplitude difference ΔA exceeds 3dB is greater than a preset number threshold, if yes, there are multiple radiation source targets, and the process goes to step S14; if no, x=x+1, and the process goes back to step S11;

[0015] S14, receiving signals s from each antenna x (t) Perform DOA estimation and obtain the angle measurement result x i , the passive radar includes a total of X-way antennas.

[0016] The passive radar anti-active decoy method based on DOA clustering algorithm described in the present invention is, as a preferred embodiment, in step S11, the value range of the preset time threshold is 10ns to 500ns.

[0017] The passive radar anti-active decoy method based on DOA clustering algorithm described in the present invention is, as a preferred embodiment, in step S13, the value range of the preset number threshold is 3 to 10.

[0018] The passive radar anti-active decoy method based on DOA clustering algorithm described in the present invention, as a preferred embodiment, step S2 includes the following steps:

[0019] S21, accumulate angle measurement results x i , and obtain the angle measurement result sample set X, X={x 1 ,x 2 ,L,x N};

[0020] S22, using an extreme value normalization method to normalize the angle measurement result sample set X to a range of 0 to 1 to obtain a normalized sample set to be sorted;

[0021] S23, selecting two sample vectors with the longest distance from the normalized sample set to be sorted, where the distance between the two sample vectors is D, and setting a clustering threshold λ according to a preset ratio threshold and the longest distance D, where the clustering threshold λ is the preset ratio threshold multiplied by the longest distance D;

[0022] S24, take the first sample in the normalized sample set to be sorted as the cluster center z of the first category 1 , will be combined with the cluster center z of the first category 1 The samples whose distance does not exceed the threshold λ are classified into the first category, and the samples with the cluster center z of the first category are 1 The samples whose distance exceeds or equals the threshold λ are taken as new cluster centers;

[0023] S25, merging the classes whose number of elements is less than the preset number threshold into adjacent classes, and obtaining all classes z 1 ,z 2 ,L z q, clustering is completed.

[0024] The passive radar anti-active decoy method based on DOA clustering algorithm described in the present invention is, as a preferred embodiment, in step S23, the value range of the preset ratio threshold is 0.1 to 0.4.

[0025] The passive radar anti-active decoy method based on DOA clustering algorithm described in the present invention is, as a preferred embodiment, in step S25, the value range of the preset quantity threshold is 2 to 5.

[0026] The passive radar anti-active decoy method based on DOA clustering algorithm described in the present invention, as a preferred embodiment, step S3 includes the following steps:

[0027] S31. Calculate all classes z 1 ,z 2 ,L z q Class Center Q ,

[0028]

[0029] S32, calculate the centroid O of the angle measurement result sample set X:

[0030] Where X = {x 1 ,x 2 ,L,x N};

[0031] S33. Calculate the mean of the dispersion of the class center. The mean of the dispersion of the class center is:

[0032] If the discreteness of the cluster center is greater than the preset discreteness threshold, the process proceeds to step S4; if not, the process returns to step S1.

[0033] The passive radar anti-active decoy method based on DOA clustering algorithm described in the present invention is, as a preferred embodiment, in step S33, the value range of the preset discreteness threshold is 3 to 6.

[0034] The passive radar anti-active decoy method based on DOA clustering algorithm described in the present invention, as a preferred embodiment, step S4 includes the following steps:

[0035] S41, extracting TOA from the clustering results, and calculating the difference between TOA in different clusters;

[0036] S42. Calculate the standard deviation of the TOA difference, and the class with the smallest standard deviation is the target angle.

[0037] The technical solution of the present invention is: a passive radar anti-active decoy method based on DOA clustering algorithm, the steps are as follows:

[0038] 1) By changing the power and phase difference of the received signal, determine whether there are multiple radiation source targets and perform DOA (direction of arrival) estimation;

[0039] 2) Over a period of time, the angle results of DOA estimation in step 1) are accumulated and DDC clustering is performed;

[0040] 3) Determine the clustering result in step 2) to see whether the dispersion of all cluster centers is greater than a threshold. If the threshold requirement is met, proceed to step 4);

[0041] 4) Perform repetition frequency stability calculation on the clustering results obtained in step 3) and output the target angle.

[0042] In step 1), the power and phase difference jump of the received signal and the DOA estimation are specifically performed through the following steps:

[0043] 1.1) Take two antennas to receive the signal s 1 (t) and s 2 (t), t=1,...,T; take the sampling point data s within the preset time threshold in the leading edge of the signal 1 (m) and s 2 (m),m=1,...,M, where M=T,s 1 (t) and s 2 (t) Take the phase difference s 1 (m) and s 2 (m) Phase difference and The difference in dephasing ΔΦ;

[0044] 1.2) Take the antenna receiving signal s 1 The average amplitude of (t) t ,s 1 The average amplitude of (m) n , A t With A n Take the amplitude difference ΔA;

[0045] 1.3) Repeat steps 2.1) to 2.2) K times. When the accumulated number of phase differences exceeding 60° and amplitude differences exceeding 3dB among the K phase differences ΔΦ and the K amplitude differences ΔA is greater than the preset number threshold, proceed to step 1.4);

[0046] 1.4) Receive signal s for each antenna x (m),m=1,...,M,x=1,...,X, where x=1,...,X is the X-path antenna, perform DOA estimation, and get the angle

[0047] In step 2), the angle results of DOA estimation are accumulated and dynamic distance clustering is performed, which is specifically performed through the following steps:

[0048] 2.1) Accumulate the angle measurement results after processing in 1.4) There are N samples X = {x 1 ,x 2 ,L,x N};

[0049] 2.2) Use extreme value normalization method to normalize the samples in 2.1) to the range of 0 to 1;

[0050] 2.3) For the normalized sample set to be sorted obtained in step 2.2), two sample vectors with the longest distance are selected, and the ratio of the clustering threshold λ to the longest distance D is the preset ratio threshold;

[0051] 2.4) For the normalized sample set to be sorted obtained in step 2.2), the first sample is used as the cluster center z 1 , the samples whose distance from the center does not exceed the threshold λ are classified into the nearest class, otherwise, they are used as new cluster centers;

[0052] 2.5) Merge the classes whose number of elements is less than the preset threshold into the adjacent classes to obtain all classes z 1 ,z 2 ,Lz q , completing the entire clustering process.

[0053] The discreteness determination of all class centers in step 3) is specifically as follows:

[0054] 3.1) For the clustering result z in 2.5) 1 ,z 2 ,L z q ,

[0055] Corresponding class center

[0056] 3.2) Calculate all samples X = {x 1 ,x 2 ,L,x N The centroid of

[0057] 3.3) Calculation The mean of is the discreteness of the class center. If it is greater than the preset discreteness threshold, go to step 4).

[0058] Perform repetition frequency stability calculation on the clustering results in step 3.3) and output the target angle.

[0059] The specific steps are as follows:

[0060] 4.1) For each class in 3.3), extract the signal arrival time (TOA) in each class and calculate the difference between TOAs in different classes.

[0061] 4.2) For the TOA difference in 3.1), calculate the standard deviation and output the class with the minimum standard deviation as the target angle.

[0062] Among them, DOA is the direction of arrival of the incoming wave, and TOA is the time of arrival of the signal.

[0063] The present invention has the following advantages:

[0064] (1) The present invention uses the power and phase difference jump of the received signal to determine whether there are multiple radiation source targets, and can accurately determine whether there is active decoy in the current environment.

[0065] (2) The solution of the present invention has high real-time performance and low algorithm complexity. It is a relatively practical passive radar anti-active decoy method, and can be used for direction finding and identification of fixed-position radiation source targets on the ground by airborne / missile-borne passive radars.

[0066] (3) Airborne / missile-borne passive radar receives ground radiation source targets and performs real-time DOA estimation and clustering. By processing the DOA clustering results, it can identify and determine the targets of the active decoy system. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] Figure 1 This is a flow chart of a passive radar anti-active decoy method based on DOA clustering algorithm;

[0068] Figure 2a A schematic diagram of the received signal amplitude of a passive radar anti-active decoy method based on DOA clustering algorithm when there is no decoy source;

[0069] Figure 2b A schematic diagram of the phase difference of the received signal when there is no decoy source in a passive radar anti-active decoy method based on DOA clustering algorithm;

[0070] Figure 3a It is a passive radar anti-active decoy method based on DOA clustering algorithm. It is the received signal amplitude and phase difference image in the presence of decoy source.

[0071] Figure 3b It is a passive radar anti-active decoy method based on DOA clustering algorithm. It is the received signal amplitude and phase difference image in the presence of decoy source.

[0072] Figure 4 This is an image of angle measurement results accumulated from multiple samples of a passive radar anti-active decoy method based on DOA clustering algorithm;

[0073] Figure 5 This is a dynamic distance clustering image of a passive radar anti-active decoy method based on DOA clustering algorithm;

[0074] Figure 6 It is a clustering result image of a passive radar anti-active decoy method based on DOA clustering algorithm, in which the discreteness of the cluster center is less than the preset discreteness threshold;

[0075] Figure 7 It is a clustering result image of a passive radar anti-active decoy method based on DOA clustering algorithm, and the discreteness of the cluster center is greater than the preset discreteness threshold;

[0076] Figure 8 This is a schematic diagram of the standard deviation calculation results of four classes of a passive radar anti-active decoy method based on DOA clustering algorithm. DETAILED DESCRIPTION

[0077] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0078] Example 1

[0079] like Figure 1 As shown, a passive radar anti-active decoy method based on DOA clustering algorithm is characterized by comprising the following steps:

[0080] S1. Determine whether there are multiple radiation source targets based on the power and phase difference jump of the passive radar receiving signal. If so, perform DOA estimation to obtain the angle measurement result x i , angle measurement result x i for And go to step S2; if not, continue to receive signals;

[0081] S11. Calculate the passive radar two-way antenna receiving signal s x (t), s x+1 Phase difference (t) And the antenna receives the signal s x (t) and s x+1 (t) Sampling point data within the preset time threshold of the front part s x (m), s x+1 Phase difference (m) make and The difference is obtained by taking the phase difference ΔΦ, where x=1,...,X, x is initialized to 1, t=1, Λ, T; m=1, Λ, M; M=T;

[0082] S12, calculate the antenna receiving signal sx The average amplitude of (t) t , sampling point data s x The average amplitude of (m) m , A t With A m Taking the difference, we get the amplitude difference ΔA;

[0083] S13, judging whether the cumulative number of times that the phase difference ΔΦ exceeds 60° and the amplitude difference ΔA exceeds 3dB is greater than a preset number threshold, if yes, there are multiple radiation source targets, and the process goes to step S14; if no, x=x+1, and the process goes back to step S11;

[0084] S14, receiving signals s from each antenna x (t) Perform DOA estimation and obtain the angle measurement result x i , the passive radar includes a total of X antennas;

[0085] The preset time threshold value ranges from 10ns to 500ns;

[0086] The preset number of times threshold value ranges from 3 to 10;

[0087] S2. Accumulate the angle measurement results to obtain the angle measurement result sample set X and perform DDC clustering to obtain the clustering result. The clustering result is all classes, and all classes are z 1 ,z 2 ,L z q ;

[0088] S21, accumulate angle measurement results x i , and obtain the angle measurement result sample set X, X={x 1 ,x 2 ,L,x N};

[0089] S22, using an extreme value normalization method to normalize the angle measurement result sample set X to a range of 0 to 1 to obtain a normalized sample set to be sorted;

[0090] S23, selecting two sample vectors with the longest distance from the normalized sample set to be sorted, the distance between the two sample vectors is D, and setting a clustering threshold λ according to a preset ratio threshold and the longest distance D, the clustering threshold λ being the preset ratio threshold multiplied by the longest distance D; the value range of the preset ratio threshold is 0.1 to 0.4;

[0091] S24, take the first sample in the normalized sample set to be sorted as the cluster center z of the first category 1 , will be combined with the cluster center z of the first category 1 The samples whose distance does not exceed the threshold λ are classified into the first category, and the samples with the cluster center z of the first category are 1The samples whose distance exceeds or equals the threshold λ are taken as new cluster centers;

[0092] S25, merging the classes whose number of elements is less than the preset number threshold into adjacent classes, and obtaining all classes z 1 ,z 2 ,L z q , clustering is completed; the value range of the preset quantity threshold is 2 to 5;

[0093] S3, determine whether the central dispersion of the class is greater than the threshold, if yes, go to step S4, if not, return to step S1; S31, calculate all class z 1 ,z 2 ,L z q Class Center Q ,

[0094]

[0095] S32, calculate the centroid O of the angle measurement result sample set X:

[0096] Where X = {x 1 ,x 2 ,L,x N};

[0097] S33. Calculate the mean of the dispersion of the class center. The mean of the dispersion of the class center is:

[0098] If the discreteness of the cluster center is greater than the preset discreteness threshold, the process proceeds to step S4; if not, the process returns to step S1; the preset discreteness threshold has a value range of 3 to 6;

[0099] S4, performing repetition frequency stability calculation on the clustering results to obtain the target angle;

[0100] S41, extracting TOA from the clustering results, and calculating the difference between TOA in different clusters;

[0101] S42. Calculate the standard deviation of the TOA difference, and the class with the smallest standard deviation is the target angle.

[0102] Example 2

[0103] like Figure 1 As shown in FIG. 1 , a passive radar anti-active decoy method based on DOA clustering algorithm is shown in FIG. 1 . The steps are as follows:

[0104] (1) By analyzing the power and phase difference jumps of the received signal, it is determined whether there are multiple radiation source targets and the DOA (direction of arrival) is estimated;

[0105] The power and phase difference jump of the received signal and the DOA estimation are specifically carried out through the following steps:

[0106] (1.1) Take two antennas to receive the signal s 1 (t) and s 2 (t), t=1,...,T; take the sampling point data s within the preset time threshold (here 10ns~500ns) in the leading edge of the signal 1 (m) and s 2 (m),m=1,...,M, where M=T,s 1 (t) and s 2 (t) Take the phase difference s 1 (m) and s 2 (m) Phase difference and The difference in dephasing ΔΦ;

[0107] (1.2) Take the antenna receiving signal s 1 The average amplitude of (t) t ,s 1 The average amplitude of (m) n , A t With A n Take the amplitude difference ΔA;

[0108] (1.3) Repeat steps (1.1) to (1.2) K times. When the number of phase differences ΔΦ and amplitude differences ΔA that exceed 60° and 3 dB is greater than a preset number threshold (here 3 to 10), proceed to step (1.4);

[0109] (1.4) The signal s received by each antenna x (m),m=1,...,M,x=1,...,X, where x=1,...,X is the X-path antenna, perform DOA estimation, and get the angle

[0110] For different ground targets, different angle resolutions, and the distance between the passive radar and the target, the thresholds are set differently. For example, for a high-resolution passive radar, the mid-frequency sampling frequency can reach about 100MHz, and a set of angle estimation results can be obtained within 10ns. Figure 2a , 2b Figure 3 shows the amplitude and phase difference of the received signal when there is no bias source; Figure 4 shows the amplitude and phase difference of the received signal when there is a bias source. By comparison, it can be clearly seen that Figure 3a , 3bAccording to this feature, the leading edge of the signal is intercepted and the time threshold is set, with a range of 10 to 500 ns.

[0111] (2) Over a period of time, the angle results of DOA estimation in step (1) are accumulated and dynamic distance clustering is performed;

[0112] Accumulate the angle results of DOA estimation and perform dynamic distance clustering, which is specifically carried out through the following steps:

[0113] (2.1) Accumulate the angle measurement results after processing in (1.4) There are N samples X = {x 1 ,x 2 ,L,x N};

[0114] (2.2) Use the extreme value normalization method to normalize the samples in (2.1) to the range of 0 to 1;

[0115] (2.3) Processing step (2.2) to obtain the normalized sample set to be sorted, select the two sample vectors with the farthest distance, and the ratio of the clustering threshold λ to the farthest distance D is the preset ratio threshold (here 0.1 to 0.4);

[0116] (2.4) For the normalized sample set to be sorted obtained after processing in step (2.2), the first sample is used as the cluster center z 1 , the samples whose distance from the cluster center is less than the threshold λ are classified into the nearest class, otherwise, they are used as new cluster centers;

[0117] (2.5) Classes with less than a preset threshold (here 2 to 5) are merged into adjacent classes to obtain all classes z. 1 ,z 2 ,L z q , completing the entire clustering process.

[0118] Figure 4 The angle measurement results accumulated by multiple samples within a period of time are clustered dynamically. The results are as follows: Figure 5 As shown, it can be clearly seen that there are 4 classes after clustering, and there are currently 4 radiation source targets at different angles.

[0119] (3) Determine whether the clustering result in step (2) is greater than a threshold for the dispersion of all cluster centers. If the threshold requirement is met, proceed to step (4).

[0120] The discreteness determination of all class centers is as follows:

[0121] (3.1) For the clustering result z in (2.5) 1 ,z2 ,L z q , corresponding to the class center

[0122] (3.2) Calculate all samples X = {x 1 ,x 2 ,L,x N The centroid of

[0123] (3.3) Calculation The mean of is the discreteness of the class center. If it is greater than the preset discreteness threshold (here 3 to 6), proceed to step (4); otherwise, repeat (2.1) to (3.3) until the discreteness of the class center is greater than the preset discreteness threshold, and proceed to step (4).

[0124] Figure 6 The clustering result is when the discreteness of the cluster center is less than the preset discreteness threshold. Since the distance between the radiation source targets is less than the angle resolution, the angles of multiple targets cannot be distinguished. Figure 7 The clustering result is when the discreteness of the cluster center is greater than the preset discreteness threshold. After the passive radar spatial position is changed, different targets can be distinguished by angle.

[0125] (4) performing repetition frequency stability calculation on the clustering results obtained in step (3) and outputting the target angle;

[0126] The target point determination is carried out through the following steps:

[0127] (4.1) For the clustering results obtained in step (3), extract the signal arrival time (TOA) in each class and calculate the difference between TOA in different classes. The corresponding calculation result is:

[0128] ΔTOA1 n =TOA1 n -TOA1 n-1 =PRI1+ω1 n +Δ R

[0129] ΔTOA2 n =TOA2 n -TOA2 n-1 =PRI2+ω2 n +Δ R

[0130] ΔTOA3 n =TOA3 n -TOA3 n-1 =PRI3+ω3 n +Δ R

[0131] ΔTOA4 n =TOA4 n -TOA4 n-1 =PRI4+ω4 n +Δ R

[0132] Where PRIi is the target repetition rate, ωi n is the error of the target PRIi, Δ R is the measurement error of the passive radar receiver;

[0133] (4.2) Calculate the standard deviation of the TOA differences of each class obtained in step (4.1), and output the class with the smallest standard deviation as the target angle.

[0134] Figure 8 Schematic diagram of the standard deviation calculation results for 4 classes, with class 2 selected as the target output.

[0135] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A passive radar anti-active decoy method based on DOA clustering algorithm, Features: The following steps are involved: S1. Determine whether there are multiple radiation source targets based on the power and phase difference jump of the passive radar receiving signal. If so, perform DOA estimation to obtain the angle measurement result x i , the angle measurement result x i for And go to step S2; if not, continue to receive signals; S2, accumulating the angle measurement results to obtain an angle measurement result sample set X and performing DDC clustering to obtain a clustering result, wherein the clustering result is all classes, and the all classes are z 1 ,z 2 ,…z q ; S3, determining whether the center dispersion of the class is greater than a threshold, if yes, proceeding to step S4, if no, returning to step S1; Step S3 includes the following steps: S31. Calculate all classes z 1 ,z 2 ,…z q Class Center Q , S32, calculating the centroid O of the angle measurement result sample set X: Where X = {x 1 ,x 2 ,…,x N }; S33, calculating the mean of the dispersion of the cluster center, the mean of the dispersion of the cluster center is: If the discreteness of the cluster center is greater than the preset discreteness threshold, proceed to step S4, if not, return to step S1; S4. Perform repetition frequency stability calculation on the clustering result to obtain a target angle.

2. According to the method for passive radar anti-active decoy based on DOA clustering algorithm described in claim 1, Features: Step S1 includes the following steps: S11, calculating the passive radar two-way antenna receiving signal s x (t), s x+1 Phase difference (t) And the antenna receives the signal s x (t) and s x+1 (t) Sampling point data within the preset time threshold of the front part s x (m), s x+1 Phase difference (m) make and The difference is obtained by taking the phase difference ΔΦ, where x = 1, ..., X, x is initialized to 1, t = 1, ..., T; m = 1, ..., M; M < <T; S12, calculating the antenna receiving signal s x The average amplitude of (t) t , the sampling point data s x The average amplitude of (m) m , A t With A m Taking the difference, we get the amplitude difference ΔΑ; S13, judging whether the cumulative number of times that the phase difference ΔΦ exceeds 60° and the amplitude difference ΔΑ exceeds 3dB is greater than a preset number threshold, if yes, there are multiple radiation source targets, and the process goes to step S14; if no, x=x+1, and the process goes back to step S11; S14, receiving signals s from each antenna x (t) Perform DOA estimation to obtain the angle measurement result x i , the passive radar includes a total of X antennas.

3. According to the method for passive radar anti-active decoy based on DOA clustering algorithm described in claim 2, Features: In step S11, the preset time threshold value ranges from 10ns to 500ns.

4. According to claim 2, a passive radar anti-active decoy method based on DOA clustering algorithm, Features: In step S13, the preset number threshold value ranges from 3 to 10.

5. According to the method for passive radar anti-active decoy based on DOA clustering algorithm described in claim 1, Features: Step S2 includes the following steps: S21, accumulating the angle measurement results x i , obtain the angle measurement result sample set X, X = {x 1 ,x 2 ,…,x N }; S22, using an extreme value normalization method to normalize the angle measurement result sample set X to a range of 0 to 1 to obtain a normalized sample set to be sorted; S23, selecting two sample vectors with the farthest distance from the normalized sample set to be sorted, where the distance between the two sample vectors is D, and setting a clustering threshold λ according to a preset ratio threshold and the farthest distance D, where the clustering threshold λ is the preset ratio threshold multiplied by the farthest distance D; S24, taking the first sample in the normalized sample set to be sorted as the cluster center z of the first category 1 , will be the cluster center z of the first category 1 The samples whose distance to the first class does not exceed the threshold λ are classified into the first class, and the samples with the cluster center z of the first class are 1 The samples whose distance exceeds or equals the threshold λ are taken as new cluster centers; S25, merging the classes whose number of elements is less than the preset number threshold into adjacent classes, and obtaining all classes z 1 ,z 2 ,…z q , clustering is completed.

6. According to claim 5, a passive radar anti-active decoy method based on DOA clustering algorithm, Features: In step S23, the preset ratio threshold value has a value range of 0.1 to 0.

4.

7. According to claim 5, a passive radar anti-active decoy method based on DOA clustering algorithm, Features: In step S25, the preset quantity threshold value ranges from 2 to 5.

8. According to the method for passive radar anti-active decoy based on DOA clustering algorithm described in claim 1, Features: In step S33, the preset discreteness threshold value ranges from 3 to 6.

9. The passive radar anti-active decoy method based on DOA clustering algorithm according to claim 1, Features: Step S4 includes the following steps: S41, extracting the TOA in the clustering result, and calculating the difference between the TOAs in different clusters; S42: Calculate the standard deviation of the TOA differences, and the class with the smallest standard deviation is the target angle.

Images